抄録
A multivariable nonlinear control procedure is formulated for the control of a binary distillation column, and its set-point tracking and regulatory performances are examined and compared with those of the quadratic dynamic matrix control (QDMC) through simulation study. Direct integration of material balances at all trays is carried out to find the predicted outputs for the different values of input in the optimization process using the successive quadratic programming.
The outcome indicates that the proposed control of this study leads to better set-point tracking and regulatory performances than those of the QDMC because the nonlinear model gives better output prediction than the step response model of the QDMC for nonlinear process. It is also proved that the complex nonlinear process model can be implemented for the real time control of multivariable systems without model simplification. In addition, a tuning guideline of the nonlinear control is suggested through stability analysis.
